Interfacial charge transfer engineering in V₃O₅/ZnO p–n heterostructures for enhanced SERS performance

Abstract

Surface-enhanced Raman scattering (SERS) substrates based on transition metal oxides have garnered significant attention due to their structural diversity, tunable properties, and unique optical characteristics. However, their practical applications have been severely hindered by relatively weak Raman enhancement effects and poorly understood charge transfer mechanisms. To address these limitations, this study developed a novel charge-transfer-enhanced SERS substrate by constructing a V₃O₅/ZnO p–n semiconductor heterojunction. The optimized heterostructure demonstrates a detection limit of 10⁻⁸ M and an enhancement factor of 5.9 × 10⁴ for methylene blue (MB), while maintaining excellent stability under extreme conditions (30–70 °C and pH 3–10). Using picosecond-resolution transient absorption spectroscopy, the charge transfer dynamics at the heterojunction interface were precisely monitored, providing systematic insights into the microscopic mechanism underlying the Raman enhancement. This study provides a strategy for designing high-performance SERS substrates based on non-precious metals and insights into the charge transfer mechanism revealed by transient absorption spectroscopy.

---

Yaxin Zhai

Yaxin Zhai is a “Xiaoxiang” Scholar Distinguished Professor in the Department of Physics at Hunan Normal University (HNNU). She received her Ph.D in Physics from the University of Utah in 2017, where she conducted research on ultrafast spectroscopy of novel semiconductors in Prof. Zeev V. Vardeny's group, and subsequently completed postdoctoral training at the National Renewable Energy Laboratory (NREL) in collaboration with Senior Researcher Fellow Dr Matthew C. Beard. In 2021, Prof. Zhai joined HNNU, where she has established an independent research group that explores ultrafast spin and charge dynamics, terahertz spectroscopy, and advanced optoelectronic device architectures, with a particular focus on hybrid semiconductors, low-dimensional materials, and chiral materials.

Xingang Zhang

Xingang Zhang is an associate professor in the Department of Physics at Hunan Normal University (HNNU). He obtained his Ph.D in 2019 from the School of Physics and Technology, Wuhan University. From 2020 to 2022, he conducted his postdoctoral research at Hunan University, where his work primarily focused on the design, characterization, and practical applications of novel semiconductor materials. His current research interests center on developing innovative surface-enhanced Raman spectroscopy (SERS) substrates and investigating the fundamental charge transfer mechanisms between these substrates and target molecules.

---

1. Introduction

Surface-enhanced Raman scattering (SERS), a rapid, non-destructive analytical technique, enables precise molecular identification through vibrational fingerprint detection. Current theoretical frameworks primarily attribute SERS signal amplification to electro-magnetic (EM) and chemical (CM) enhancement mechanisms.^1,2 EM originates from surface plasmon resonance (SPR) induced by laser excitation on nanostructured noble metals, generating localized electromagnetic field amplification that elevates molecular Raman signals by 10⁴–10¹¹-fold, constituting the predominant mechanism in SERS.^3,4 However, SERS substrates based on noble metals not only face disadvantages such as high manufacturing costs, poor biocompatibility, strong spectral background, and low repeatability, but also lack selectivity for molecular detection, which seriously limits their practical applications.⁵ CM arises from molecule–substrate charge transfer interactions that elevate molecular polarization, thereby augmenting the Raman scattering cross section through altered electronic states—a mechanism distinct from electromagnetic effects.^6,7 Semiconductor materials have become the main research focus for providing CM due to their diverse types and highly adjustable band structures.^8–11 CM based SERS substrates have shown significant advantages in biocompatibility and selective detection, but the low Raman enhancement ability has always been a key limiting factor for their applications.

In order to prepare high-performance CM based SERS substrates, researchers have extensively studied materials such as semiconductor oxides,¹² transition metal sulfides,^13,14 two-dimensional materials,^15,16 MXenes,^17,18etc., exploring their potential as high-performance SERS substrates. Research has found that defects,^19,20 element ratios,^21,22 element doping,^23,24 and crystal phase^25,26 in materials can greatly affect their Raman enhancement ability. Rational modulation of material parameters enables significant optimization of substrate–molecule charge-transfer processes, thereby amplifying molecular Raman responses. This principle is exemplified in surface defect engineering strategies, as demonstrated by Cong et al.‘s work, where controlled oxygen vacancies in ultrathin WO₃ nanosheets elevated charge-carrier density, triggering interfacial electronic coupling effects that achieved ∼100-fold Raman signal amplification.²⁷ Xu et al. constructed SnSe_2−x nanoflake arrays; by adjusting the ratio of Sn and Se, SnSe_1.75 markedly improves R6G's Raman spectral response, primarily due to selenium vacancy formation.²⁸ Jiang et al. optimized SnO₂'s band structure by varying Ce dopant concentration, improving charge transfer between the substrate and target molecule, which led to enhanced Raman signals of the molecule.²⁹ Meng et al. used Xe lamp irradiation to transform crystalline MoO₃ into an amorphous state and found that amorphous MoO₃ has a higher Raman enhancement ability. This enhancement is attributed to its higher electron delocalization and electronic state density, which facilitate electron transfer between the substrate and target molecules.³⁰

Beyond modifying individual materials, constructing heterojunctions effectively enhances the performance of non-noble-metal SERS substrates. Zhou et al. demonstrated that sponge-like Cu-doped SnO₂–NiO p–n heterostructures show superior Raman enhancement, resulting from the combined effects of p–n junction-enhanced charge separation efficiency and Cu doping-induced charge transfer resonance.³¹ Xie et al. developed TiO₂@MoO_x nanorod heterojunctions for SERS, achieving a 10⁻⁸ M detection limit for R6G. Their mechanistic study revealed that the staggered band alignment in TiO₂@MoO_x enables photo-generated electron–hole separation in the depletion layer under laser excitation. Electrons migrate to MoO_x surfaces and subsequently transfer to adsorbed molecules, amplifying SERS signals.³² This demonstrates that optimizing interfacial charge transfer is critical for enhancing molecular Raman detection.

Herein, we assemble Zn-based metal organic framework (MOF) nanoparticles and V-based MOF nanowires into hybrid MOF-on-MOF heterostructures, which are further annealed to transform into V₃O₅/ZnO p–n semiconductor heterostructures as an effective charge transfer SERS substrate. Experimentally, methylene blue (MB) served as a Raman probe to assess the SERS activity of materials. Comparative tests under 532 nm laser excitation revealed that the V₃O₅/ZnO heterojunction outperformed standalone V₃O₅ nanowires and ZnO nanoparticles, achieving an MB enhancement factor of 5.9 × 10⁴ and a detection limit of 10⁻⁸ M. Transient absorption (TA) spectroscopy confirmed that the heterojunction's superior SERS performance stems from intensified charge-transfer interactions with MB, which amplify molecular polarization and boost Raman signals.

2. Results and discussion

2.1 Preparation of V₃O₅/ZnO heterojunctions

The preparation process of V₃O₅/ZnO heterojunctions is shown in Fig. 1. Firstly, V-MOF nanowires with uniform structure were synthesized by the hydrothermal method, and then the Zn-MOF was grown on the surface of V-MOF nanowires to construct tomatoes-on-sticks shaped hybrid MOF-on-MOF heterostructures. Finally, after annealing treatment, the hybrid MOF-on-MOF heterostructures were converted into V₃O₅/ZnO heterojunctions.

Fig. 1 Schematic illustration of the synthesis of V₃O₅/ZnO heterojunctions.

As shown in Fig. 2a, V-MOF nanowires exhibit uniform surfaces with an average diameter of ∼210 nm via scanning electron microscopy (SEM) analysis. Transmission electron microscopy (TEM) imaging (Fig. 2b and c) corroborates their structural uniformity, aligning with SEM observations. X-ray diffraction (XRD) patterns (Fig. S1) reveal high crystallinity through sharp peaks matching prior reports.³³ These V-MOF nanowires were subsequently employed as host frameworks to construct MOF-on-MOF heterostructures. By directly immersing V-MOF nanowires into a methanol solution of 2-methyllimidazole and Zn²⁺, a V-MOF/Zn-MOF heterostructure was synthesized. As shown in Fig. 2d, each V-MOF nanowire is strung with a number of regular dodecahedron Zn-MOF particles with a size of 400–500 nm, forming a shape similar to the tomatoes-on-sticks string. With the increase of V-MOF nanowire feeding (Fig. S2), the number of Zn-MOF particles on its surface can also be well controlled. The Zn-MOF nanoparticles were synthesized as shown in Fig. S3, presenting a regular dodecahedron. The V-MOF/Zn-MOF heterostructure morphology was confirmed by TEM (Fig. 2e) and EDS mapping (Fig. 2f). XRD analysis (Fig. S4) verified its crystallinity, with all diffraction peaks corresponding to pristine V-MOF and Zn-MOF phases.^33,34 The compositions of V-MOF, Zn-MOF, and V-MOF/Zn-MOF were further validated by Raman spectroscopy (Fig. S5), with results consistent with previously reported literature.^35–37

Fig. 2 (a) SEM and (b and c) TEM images of V-MOF; (d) SEM, (e) TEM images, and (f) EDS mapping of the V-MOF/Zn-MOF heterostructure.

Through annealing treatment (300 to 500 °C), the V-MOF/Zn-MOF heterostructure is converted into a V₃O₅/ZnO nanowire heterostructure. After annealing at 600 °C, the V₃O₅/ZnO nanowires transform into agglomerated nanoparticles (Fig. S6). Moreover, V-MOF and Zn-MOF are transformed into V₃O₅ (Fig. S7) and ZnO (Fig. S8), respectively, after annealing at 500 °C, while still maintaining their nanowire and nanoparticle structures. The thermal stability of the as-obtained V-MOF/Zn-MOF (annealed at 500 °C) was investigated by thermogravimetric analysis (TGA) (Fig. S9). The results revealed two distinct degradation stages: the first stage (0–100 min) corresponds to the volatilization of solvent molecules (H₂O), while the second stage (100–300 min) is attributed to the decomposition of organic linkers and the MOF structure.³⁸ The TEM image indicates that V₃O₅/ZnO nanowires exhibit a rough surface (Fig. 3a). In the HRTEM image (Fig. 3b), the (002) and (110) crystal planes of ZnO and the (310) crystal plane of V₃O₅ can be seen, both of which have good crystallinity. The clear boundary indicates the synthesis of the V₃O₅/ZnO heterojunction. The elemental mapping images of V₃O₅/ZnO nanowires confirm its hybrid structure (Fig. 3c). The concentrations of Zn, V, and O in V₃O₅/ZnO are estimated to be 38.7 wt%, 35.9 wt%, and 19.4 wt%, respectively, via a quantitative calculation of the EDX spectrum (Fig. S10). XRD spectra of V₃O₅/ZnO nanowires were also collected, as shown in Fig. 3d. Specifically, most of the sharp diffraction peaks are attributed to V₃O₅ (PDF#71-0039), indicating that V₃O₅ has good crystallinity. Meanwhile, we also observed diffraction peaks of ZnO, which further confirm the synthesis of the V₃O₅/ZnO heterojunction, consistent with the HRTEM image. XRD analysis of the V-MOF product after high-temperature annealing reveals a composition of V₃O₅ and V₂O₃, while the Zn-MOF-derived product is a composite of ZnO and MOF, indicating that 500 °C annealing is insufficient for complete pyrolysis of pure Zn-MOFs (Fig. S11). X-ray photoelectron spectroscopy (XPS) analysis of the V₃O₅/ZnO heterojunction (Fig. S12) confirmed the presence of Zn, V, O, and C. The Zn 2p XPS spectrum of the sample exhibits a typical spin–orbit splitting with a 2p_3/2 peak (blue) at 1020.9 eV and a 2p_1/2 peak (green) at 1044.2 eV (Fig. 3e). The corresponding splitting energy of ∼23 eV is consistent with the inherent electronic characteristics of zinc. Here, the 2p_3/2 binding energy aligns closely with that of Zn²⁺ (e.g., ZnO, ∼1021.3 eV) and is significantly lower than that of metallic Zn⁰ (1021.7 eV), confirming that Zn is in the +2 oxidation state. Additionally, the weak and broad loss feature observed (rather than a sharp plasmon peak) is consistent with the electronic structure of ZnO semiconductor materials, thus providing auxiliary evidence for the existence of Zn²⁺.³⁹ XPS characterization of the V 2p spectrum (Fig. 3f) reveals that vanadium atoms in the material exhibit mixed valence states of V⁴⁺ and V³⁺. V⁴⁺ is identified as the dominant species, with its 2p_3/2 peak centered at 516.4 eV and its 2p_1/2 peak at 523.9 eV, corresponding to a spin–orbit splitting energy of approximately 7.5 eV. These spectral features are consistent with the typical characteristics of tetravalent vanadium, as exemplified by VO₂. V³⁺ exists as a minor component, showing 2p_3/2 and 2p_1/2 peaks at 515.6 eV and 522.6 eV, respectively, with a splitting energy of ∼7.0 eV, matching well with the standard binding energies of trivalent vanadium compounds such as V₂O₃. The estimated area ratio of V⁴⁺ to V³⁺ peaks is 2.07 : 1, suggesting that the material may consist of a mixed vanadium oxide.⁴⁰ Notably, compared with the binding energy values of the corresponding oxidation states in the literature,^39,41 the binding energy peak of the Zn oxidation state in the heterojunction shows a negative shift of approximately −0.4 eV, while the binding energy peak of the V oxidation state exhibits a slight positive shift (+0.1 eV). These results indicate that the electron density around Zn increases, whereas the electron density around V decreases. This difference in charge distribution promotes the formation of a built-in electric field at the V₃O₅–ZnO interface, which is directed from V₃O₅ to ZnO. The O 1s peak exhibits a multi-component structure (Fig. S13). It primarily consists of two well-assigned chemical states: the peak at 529.4 eV is attributed to the Zn–O bond (lattice oxygen in zinc oxide), while the peak at 530.7 eV corresponds to the V–O bond (lattice oxygen in vanadium oxide). Additionally, a weak, unlabeled signal is observed around 531.9 eV, which may originate from minor oxygen species, such as surface-adsorbed hydroxyl groups (–OH) or oxygen defects.³⁹ These findings collectively demonstrate the successful fabrication of V₃O₅/ZnO nanowire heterostructures.

Fig. 3 (a) TEM and (b) HRTEM images of V₃O₅/ZnO the heterostructure; (c) EDX mapping images of V, Zn, O and C element on the V₃O₅/ZnO heterostructure; (d) XRD pattern of V₃O₅/ZnO; high-resolution XPS spectra of (e) Zn 2p and (f) V 2p of the V₃O₅/ZnO heterostructure.

2.2 SERS performance of V₃O₅/ZnO heterojunctions

To assess SERS performance, MB was utilized as the probe molecule. In a typical experiment, V₃O₅/ZnO, V₃O_5, and ZnO were immersed in a 10⁻⁵ M MB aqueous solution for 2 hours, and the corresponding Raman spectra were collected under 532 nm laser excitation. Fig. 4a presents the Raman spectra of MB molecules collected from each sample, and these peak positions are consistent with literature reports.^16,22 Meanwhile, the V₃O₅/ZnO heterojunction has the best Raman enhancement performance, indicating that the heterojunction interaction between V₃O₅ and ZnO has a potential role in enhancing the Raman spectra of MB molecules. Moreover, we also compared the effect of annealing temperature on the Raman enhancement ability of the V₃O₅/ZnO heterojunction. As shown in Fig. S14, as the temperature increases from 300 °C to 500 °C, the Raman enhancement ability of the V₃O₅/ZnO heterojunction gradually improves, which can be attributed to the better crystallinity of the V₃O₅/ZnO heterojunction at higher temperatures. Thus, we chose samples annealed at 500 °C for subsequent experiments. Subsequently, MB aqueous solutions of different concentrations (10⁻⁵–10⁻⁹ M) were used to evaluate the detection ability of the V₃O₅/ZnO heterojunction. From Fig. 4b, it can be seen that even at an MB concentration as low as 10⁻⁸ M, the Raman signal can still be resolved. Even on fish surfaces, MB with 10⁻⁷ M can also be detected (Fig. S15). Based on the above Raman measurement data, we calculated the enhancement factor (EF) of the V₃O₅/ZnO heterojunction to be 5.9 × 10⁴ when the concentration of MB is 10⁻⁸ M (see the Experimental section). Moreover, to investigate the SERS enhancement capability for other molecules, we detected R6G, malachite green, and methyl orange (common organic dye pollutants) using V₃O₅/ZnO. The results showed clear Raman peaks even at concentrations as low as 10⁻⁶ M (Fig. S16), indicating that the V₃O₅/ZnO heterojunction's SERS effect is not limited to MB molecules. The above results indicate that the V₃O₅/ZnO heterojunction has superior Raman enhancement ability for MB molecules.

Fig. 4 (a) SERS spectra of MB molecules on ZnO, V₃O₅ and V₃O₅/ZnO heterojunctions; (b) SERS spectra of different MB concentrations on V₃O₅/ZnO heterojunctions; (c) the repeated intensity data of the MB 1623 cm⁻¹ peak proves the reproducibility of the V₃O₅/ZnO heterojunction; the stability of MB molecules detected by the V₃O₅/ZnO heterojunction at different (d) pH values and (e) temperatures; (f) UV-vis absorption spectra of pure MB, bare V₃O₅/ZnO heterojunctions and V₃O₅/ZnO/MB; Tauc plot of bare (g) ZnO and (h) V₃O₅; (i) the corresponding Mott–Schottky curves and schematic diagram of the energy band structure.

Raman spectra collected from ∼20 random points on the V₃O₅/ZnO heterojunction were analysed by calculating the standard deviation of the 1623 cm⁻¹ peak intensity, which showed minimal variation (RSD = 4.4%, Fig. 4c), demonstrating uniform SERS activity and homogeneous MB molecule adsorption across the heterojunction. Chemical stability is also an important indicator of SERS substrates, and SERS substrates with good chemical stability can function normally under extreme conditions. Therefore, we evaluated the detection ability of the V₃O₅/ZnO heterojunction for MB molecules at different pH values (from 3 to 10) and temperatures (30–70 °C). The Raman spectra of MB molecules showed almost no change (Fig. 4d and e). The above results indicate that the V₃O₅/ZnO heterojunction SERS substrate has potential applications under extreme conditions.

2.3 Enhancement mechanism

To reveal the Raman enhancement mechanism of the V₃O₅/ZnO heterojunction, electrochemical impedance spectroscopy (EIS) was employed to investigate the resistive behaviour of samples. The Nyquist plots of all samples exhibited semicircles in the high-frequency region and inclined straight lines in the low-frequency region (Fig. S17). The semicircle diameter reflects the charge transfer resistance at the electrolyte/electrode interface, with larger diameters indicating higher resistance. The linear portion corresponds to Warburg diffusion resistance, representing ion diffusion rates. Comparative analysis of the EIS spectra reveals that V₃O₅/ZnO demonstrates the highest conductivity. The superior electrical conductivity of V₃O₅/ZnO facilitates efficient electron transfer between the sample and MB molecules. Then, we conducted N₂ adsorption–desorption isotherm measurements and analyzed the pore-size distribution of the samples (Fig. S18). The test results demonstrate that the Brunauer–Emmett–Teller (BET) surface areas of V₃O₅, ZnO, and V₃O₅/ZnO are 99.01 m² g⁻¹, 1144.08 m² g⁻¹, and 120.15 m² g⁻¹, respectively, with the average pore size primarily distributed in the 2–10 nm range. The ZnO nanoparticles inherit the porous structure of Zn-MOF particles, exhibiting a significantly larger specific surface area. Compared to the individual V₃O₅, upon combining ZnO with V₃O₅, the specific surface area is markedly enhanced (from 99.01 m² g⁻¹ to 120.15 m² g⁻¹), thereby providing more adsorption sites for target molecules. Additionally, the larger pore size ensures efficient accessibility of target molecules to the active sites within the material. Moreover, ultraviolet-visible (UV-vis) absorption spectra were recorded to investigate the interactions between MB and the V₃O₅/ZnO heterojunction (Fig. 4f). As we can see, the MB molecule exhibits a distinct absorption peak at 655 nm, accompanied by an absorption shoulder at 615 nm, which is consistent with the literature report.²⁷ Due to the synergistic effects, the V₃O₅/ZnO heterostructure presents significant broadband absorption in the ultraviolet to near-infrared spectral region. It is noteworthy that upon adsorption of MB onto the heterojunction, the corresponding characteristic absorption peak undergoes significant broadening. This observation suggests that effective electron transfer occurs between the molecule and the semiconductor. Moreover, we compared the adsorption of MB molecules by V₃O₅, ZnO, and V₃O₅/ZnO composites using UV-vis absorption spectroscopy. Equal masses of each material were separately added to equal volumes of 10⁻⁵ M MB solutions, and the mixtures were stirred at room temperature for 1 h. After centrifugation, the supernatant was collected and analyzed by UV-vis absorption spectroscopy (Fig. S19). The results revealed that V₃O₅ exhibited the weakest MB adsorption effect, while V₃O₅/ZnO exhibited comparable adsorption efficiency to ZnO, despite ZnO's significantly higher specific surface area compared to V₃O₅/ZnO (1144.08 m² g⁻¹vs. 99.01 m² g⁻¹). This indicates that V₃O₅/ZnO has superior adsorption properties for MB molecules. To further understand the possible electron transition mechanism in the heterostructure, the corresponding Tauc plots of bare ZnO and V₃O₅ nanostructures prepared under the same conditions are displayed in Fig. 4g and h. Taking into account the Tauc eqn (1),⁴²

(αhν)^1/γ = A(hν − E_g) (1)

where α is the absorption coefficient, hν is the energy of incident light, A is the proportionality constant, and γ = 1/2 or 2 for direct or indirect band gaps, respectively. The band gap energy E_g can be evaluated by fitting the linear region of the plot to zero. Here, if a direct transition is assumed, the calculated bandgaps of ZnO and V₃O₅ are approximately 2.92 eV and 1.67 eV, respectively. This indicates that the electrons in V₃O₅ more readily absorb photons and become excited to the excited state under 532 nm light irradiation. In addition, Mott–Schottky diagrams of the corresponding V₃O₅ and ZnO electrodes are displayed in Fig. 4i. From the x-intercepts of linear extrapolations, the flat-band potentials V_fb of V₃O₅ and ZnO can be determined to be 0.84 V vs. SHE and 0.46 V vs. SHE, respectively. According to the following correlation between flat-band potential and the Fermi level (2)⁴³

E_F = −V_fb(vs. SHE)·e − 4.4 eV (2)

The corresponding E_F values for V₃O₅ and ZnO are determined to be −5.24 eV and −4.86 eV, respectively. Considering that the Fermi level of an n-type semiconductor (p-type semiconductor) is slightly lower (higher) than its conduction band bottom (valence band top) by approximately 0.3 eV, combined with the energy band gap obtained above, we can obtain the energy level arrangement of the heterostructure. A typical staggered-gap heterojunction of V₃O₅/ZnO is depicted in the inset of Fig. 4i, which facilitates the rapid separation of interface photogenerated carriers. Moreover, the bare V₃O₅ and ZnO present positive and negative slopes in the Mott–Schottky plot, respectively, indicating that V₃O₅ is a p-type semiconductor, while ZnO is an n-type semiconductor. Therefore, when p-type V₃O₅ encounters n-type ZnO, the majority carriers (electrons) in ZnO will migrate to V₃O₅ to obtain a Fermi level equilibrium. As a result, an internal electric field from ZnO toward V₃O₅ will be established, which will further accelerate the rapid extraction of photoelectrons at the interface, providing an advantageous platform for molecular Raman detection.

Furthermore, to uncover the charge transfer pathways in more depth, we investigated the charge transfer processes in heterostructures by transient absorption (TA) spectroscopy. Fig. 5a compares the TA spectra at 1 ps for thin films of V₃O₅, MB, and the V₃O₅/MB heterostructure excited at 3.1 eV. The V₃O₅ spectrum shows three features: a negative ground-state photo-bleaching (PB) at around 1.7 eV, and two positive peaks at 2.1 eV and higher energies, indicating photo-induced absorption (PIA) from the first excited state to higher states, see Fig. 5b. The MB thin film exhibits a strong negative PB at 2.1 eV and a weak PIA peak at higher energy. In the V₃O₅/MB heterostructure, the PB and PIA from MB are significantly enhanced, while those from V₃O₅ are suppressed, indicating efficient charge transfer from V₃O₅ to MB within our time resolution (∼100 fs). Fig. 5c shows the time evolution of the TA signal in the V₃O₅/MB heterostructure at 1 ps, 2 ps, and 10 ps. Fig. 5d compares the TA spectra at 1 ps for V₃O₅, MB, ZnO, and the V₃O₅/ZnO/MB heterostructure. Due to its UV-range bandgap, the ZnO thin film shows only a broad PIA peak centered at 2.3 eV. In the V₃O₅/ZnO/MB heterostructure, the PIA peaks of ZnO are significantly enhanced, while the PB and PIA signals from both V₃O₅ and MB are weakened, indicating charge transfer from V₃O₅ and MB to ZnO (Fig. 5e). Unlike the rapid charge transfer in the V₃O₅/MB heterostructure, the PIA signal in the V₃O₅/ZnO/MB heterostructure increases over the first 10 ps, indicating relatively slower charge transfer, as shown in Fig. 5f. To confirm these findings, we measured the TA spectra for V₃O₅, MB, ZnO, and their mixed materials in solutions, as shown in Fig. S20. Due to weak molecular interactions, the spectra of the mixed solutions are simply a combination of the components, with no charge-transfer processes observed. The TA spectra indicate significant charge transfer between the V₃O₅/ZnO heterostructure and the MB molecule, effectively amplifying the molecular polarizability.

Fig. 5 Time-resolved spectroscopy study of V₃O₅, MB, ZnO thin films and heterostructures. (a) TA spectra at 1 ps for V₃O₅, MB, and the V₃O₅/MB heterostructure; (b) the energy level and charge transfer process in the V₃O₅/MB heterostructure; (c) TA spectra for the V₃O₅/MB heterostructure at 1 ps, 2 ps, and 10 ps; (d) TA spectra at 1 ps for V₃O₅, MB, ZnO, and the V₃O₅/ZnO/MB heterostructure; (e) the energy level and charge transfer process in the V₃O₅/ZnO/MB heterostructure; (f) TA spectra for the V₃O₅/ZnO/MB heterostructure at 1 ps, 2 ps, and 10 ps.

3. Conclusions

In summary, V₃O₅/ZnO heterostructures derived from hybrid MOF-on-MOF heterostructures were prepared. As a Raman enhanced sensor, a detection limit of 1 × 10⁻⁸ M and an EF of 5.9 × 10⁴ for MB molecules were achieved on the V₃O₅/ZnO heterostructure. Meanwhile, the sensor can work normally under strong acidic and alkaline conditions and at high temperatures, showing a wide range of potential applications. The superior SERS properties of the V₃O₅/ZnO heterostructure are attributed to the chemical enhancement. The charge transfer mechanism between the V₃O₅/ZnO heterostructure and MB molecules was investigated in detail by TA spectroscopy, and the charge transfer process was observed on the picosecond scale. This study provides strategies for the design of high-performance chemically enhanced Raman substrates, as well as insights into the charge transfer mechanism revealed by TA spectroscopy.

4. Experimental section

4.1 Preparation of the V-MOF

The MOF synthesis commenced with preparing an aqueous precursor solution (4 mM) containing stoichiometric VCl₃ (Sinopharm Chemical Reagent Co., Ltd) and terephthalic acid under vigorous stirring (30 min, ambient conditions). Subsequent addition of CTAB (Sinopharm Chemical Reagent Co., Ltd) as a structure-directing agent preceded hydrothermal treatment (200 °C, 48 h) in a sealed autoclave. Post-synthesis processing involved (i) cooling to ambient temperature, (ii) triple-phase washing (deionized water/ethanol cycles), and (iii) vacuum drying at 60 °C for 12 h.

4.2 Preparation of the Zn-MOF

The Zn-MOF was synthesized via ambient-temperature coprecipitation. Stoichiometric Zn·(NO₃)₂·6H₂O (0.024 M) (Sinopharm Chemical Reagent Co., Ltd) and 2-methylimidazole (2.4 M) (Sinopharm Chemical Reagent Co., Ltd) aqueous solutions were rapidly mixed under vigorous stirring (3 h, 25 °C), yielding a milky suspension. This suspension was centrifuged at 5000 rpm for 5 minutes to separate the precipitate and supernatant. The supernatant was removed and the precipitate was collected. The precipitate was washed three times with methanol solution to remove impurities. Finally, the cleaned precipitate was dried at 60 °C for 12 hours to obtain white Zn MOF powder.

4.3 Preparation of the V-MOF/Zn-MOF heterostructure

A certain amount of dried V-MOF was weighed and dispersed in 4.5 mL of methanol. Meanwhile, 138.5 mg of Zn·(NO₃)₂·6H₂O and 154 mg of 2-methylimidazole were dissolved in 4 mL of ethanol solution, respectively. V-MOF solution was added to 2-methylimidazole solution; then the mixed solution was placed on a magnetic stirrer and stirred at 35 rpm for 5 minutes at room temperature to ensure uniform mixing. Afterwards, the Zn·(NO₃)₂·6H₂O solution was quickly added to the mixed solution and stirring was continued at the same speed for 3 hours until the solution turns light green. The solution was centrifuged at 5000 rpm for 5 minutes and then redispersed in 8 mL of ethanol solution; this process was repeated three times, and the precipitate was collected. Finally, the obtained precipitate was dried in a drying oven at 60 °C for 12 hours. Finally, a light green V-MOF/Zn-MOF powder was obtained.

4.4 Preparation of the V₃O₅/ZnO heterostructure

The V-MOF/Zn-MOF heterostructure was placed into a tube furnace, and argon was allowed to flow at room temperature for half an hour; then the high-temperature box furnace was heated up to the required temperature (300–600 °C) at a rate of 5 °C min⁻¹, and maintained at that temperature for 3 hours under an Ar atmosphere. After cooling naturally to room temperature, the brown V₃O₅/ZnO powder was collected. The preparation methods for V₃O₅ and ZnO powders is consistent with the procedure described above.

4.5 Sample characterization

The morphology and microstructure of samples were characterized using a field-emission transmission scanning electron microscope (SEM450) coupled with an energy-dispersive spectrometer (EDS) and a transmission electron microscope (TEM, Tecnai G2 F20 S-TWIN). The crystal structure and surface chemistry of the samples were characterized by X-ray diffraction (XRD, AXS, D8 Discover) and X-ray photoelectron spectroscopy (XPS, ESCALAB250Xi). Time-resolved spectroscopy was performed on an ultrafast transient absorption spectroscopy system (Femto-TA100). SERS measurements were performed on a Horiba LabRAM HR Evolution, with a 532 nm laser focused through a 100× objective onto the sample. The integration time for SERS measurements was 10 s.

4.6 SERS performance evaluation methods

The SERS performance of the substrate was comprehensively evaluated through four key metrics: (1) detection sensitivity (quantified by the limit of detection), (2) enhancement factor (calculated via the ratio of SERS to normal Raman signals), (3) spatial uniformity (assessed via the standard deviation of Raman spectral intensity), and (4) stability under extreme conditions (including exposure to strong acids/bases and elevated temperatures).

4.7 Calculation of Raman enhancement factor (EF)

The Raman enhancement factor (EF) of the V₃O₅/ZnO heterostructure was calculated through the following equation:

(I_SERS/N_SERS)/(I_bulk/N_bulk) (3)

where I_SERS and N_SERS are the Raman intensities and the amounts of molecules involved in the SERS experiments, and I_bulk and N_bulk are the Raman intensities and the amounts of molecules involved in the bulk Raman measurements. Here, the peak of MB at 1623 cm⁻¹ was used to estimate the EF. The peak intensity at 1623 cm⁻¹ for MB/V₃O₅/ZnO (1 × 10⁻⁸ M) is 237 counts, and that of bulk MB (1 × 10⁻² M) is 467 counts.

In the SERS experiments, 20 µL of MB solution was drop-cast on a 1 cm² SERS material followed by gentle drying. N_SERS is given by N_SERS = cVN_AA_Raman/A_sub, where c is the MB concentration, V is the MB droplet volume, N_A is the Avogadro constant, A_Raman is the laser spot area, and A_sub is the substrate area. For N_bulk, 0.1 M MB solution was drop-cast onto bare glass. After gentle drying, bulk MB crystals were formed. The density of MB was used to calculate the number of molecules. N_bulk is given by N_SERS = ρhN_AA_Raman/M, where ρ is the density of MB (1.757 g cm⁻³); the laser penetration depth h can be calculated using the equation h = 2λ/N_A², where N_A is the numerical aperture (0.5). M is the molar mass of MB (319.86 g mol⁻¹). Taking all the above mentioned factors into account, the EF in the MB/V₃O₅/ZnO measurements can be derived as:

4.8 Transient absorption (TA) measurements

Transient absorption (TA) measurements were conducted using a pump-probe spectrometer setup. Initially, a Ti: sapphire amplifier generated a fundamental laser pulse with a wavelength of 800 nm, operating at a repetition rate of 1 kHz. This fundamental pulse was then split into two branches by a beam splitter. One branch was directed toward a β-barium borate (BBO) crystal for second harmonic generation of the 800 nm pulse to generate an excitation pump at 400 nm. The pump pulse, modulated at a frequency of 500 Hz, underwent attenuation via neutral-density filter wheels. Simultaneously, the other branch of the fundamental pulse was focused into a sapphire crystal to produce a white-light continuum spanning from 350 nm to 1600 nm, utilized as the probe. Time delays between the pump and probe pulses were achieved using a motorized translation stage with a retro-reflecting mirror. TA measurements in the thin films were performed under a microscope. The pump and probe beams were combined using a 425 nm long-pass dichroic mirror, which transmits probe beams with wavelengths longer than 450 nm while reflecting pump beams with wavelengths of 380–410 nm. In this study, the probe pulse spans from 450 nm to 800 nm. Then the combined beams were directed towards the microscope objective (Olympus, 50×magnification, numerical aperture 0.75, and working distance 3 mm). The pump and probe beams were spatially overlapped on the sample surface, both incident on the sample normally. The focused spot size of the pump beam on the sample surface was approximately 1 µm. The transmitted probe beam of the sample was focused onto another 50× objective and collected through a fiber optic spectrometer.

Author contributions

Zhang X., Liang S., and Zhai Y. conceived the concept, designed the research, and wrote the manuscript. Wu X. and Yang C. performed the experiments. Liu Y., Wang J. and Zhai Y. conducted transient absorption spectroscopy measurements and analysis. All authors contributed to the general discussion.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the electronic supplementary information (SI). No additional datasets or code were generated or analysed in this study. Supplementary information: SEM images and XRD patterns; Raman spectra and TG/DTG curves of V-MOF/Zn-MOF; EDX spectra and XPS spectrum of V₃O₅/ZnO heterostructure; SERS spectra of molecules detected on V₃O₅/ZnO heterostructure; EIS curves, N₂ adsorption–desorption isotherms and the pore-size distribution of V₃O₅, ZnO, and V₃O₅/ZnO; UV-Vis absorption spectra and time-resolved spectroscopy. See DOI: https://doi.org/10.1039/d5ta04365d.

Acknowledgements

This research was supported by the Science & Technology Department (Grant No. 2024JJ5250, 2023ZJ1010 and 2024JJ2041), the Education Department of Hunan Province (Grant No. 23A0047), and the National Natural Science Foundation of China (NSFC) (Grant No. 12204167, 62522404, and 12421005).

References

1. T. Itoh, M. Procházka, Z. Dong, W. Ji, Y. Yamamoto, Y. Zhang and Y. Ozaki, Chem. Rev., 2023, 12, 1552–1634.
2. H. Ma, S. Pan, W. Wang, X. Yue, X. Xi, S. Yan, D. Wu, X. Wang, G. Liu and B. Ren, ACS Nano, 2024, 18, 14000–14019.
3. E. Ru and B. Auguié, ACS Nano, 2024, 18, 9773–9783.
4. S. Ding, J. Yi, J. Li, B. Ren, D. Wu, R. Panneerselvam and Z. Tian, Nat. Rev. Mater., 2016, 1, 1–16.
5. X. Liu, Z. Ye, Q. Xiang, X. Liu, Z. Ye, Q. Xiang, Z. Xu, W. Yue, C. Li, Y. Xu, L. Wang, X. Cao and J. Zhang, Chem, 2023, 9, 1464–1476.
6. I. Chaudhry, G. Hu, H. Ye and L. Jensen, ACS Nano, 2024, 18, 20835–20850.
7. S. Cong, X. Liu, Y. Jiang, W. Zhang and Z. Zhao, Innovation, 2020, 1, 100051.
8. Q. Wang, K. Chang, Q. Yang and W. Wu, Trends Food Sci. Technol., 2024, 104460.
9. S. Li, H. Rao, C. Tang, T. Bao, P. She and J. Qin, Coord. Chem. Rev., 2024, 510, 215803.
10. D. Cialla-May, A. Bonifacio, T. Bocklitz, A. Markin, N. Markina, S. Fornasaro, A. Dwivedi, T. Dib, E. Farnesi, C. Liu, A. Ghosh and J. Popp, Chem. Soc. Rev., 2024, 53, 8957–8979.
11. X. Tang, Q. Hao, X. Hou, L. Lan, M. Li, L. Yao, X. Zhao, Z. Ni, X. Fan and T. Qiu, Adv. Mater., 2024, 2312348.
12. S. Cong, Z. Wang, W. Gong, Z. Chen, W. Lu, J. Lombardi and Z. Zhao, Nat. Commun., 2019, 10, 678.
13. Z. Li, L. Zhai, Q. Zhang, W. Zhai, P. Li, B. Chen, C. Chen, Y. Yao, Y. Ge, H. Yang, P. Qiao, J. Kang, Z. Shi, A. Zhang, H. Wang, J. Liang, J. Liu, Z. Guan, L. Liao, V. Neacsu, C. Ma, Y. Chen, Y. Zhu, C. Lee, L. Ma, Y. Du, L. Gu, J. Li, Z. Tian, F. Ding and H. Zhang, Nat. Mater., 2024, 23, 1–8.
14. Y. Peng, C. Lin, Y. Li, Y. Gao, J. Wang, J. He, Z. Huang, J. Liu, X. Luo and Y. Yang, Matter, 2022, 5, 694–709.
15. J. Jin, Z. Guo, D. Fan and B. Zhao, Mater. Horiz., 2023, 10, 1087–1104.
16. Q. Lv, J. Tan, Z. Wang, P. Gu, H. Liu, L. Yu, Y. Wei, L. Gan, B. Liu, J. Li, F. Kang, H. Cheng, Q. Xiong and R. Lv, Nat. Commun., 2023, 14, 2717.
17. Z. He, T. Rong, Y. Li, J. Ma, Q. Li, F. Wu, Y. Wang and F. Wang, ACS Nano, 2022, 16, 4072–4083.
18. G. Yang, F. Liu, J. Zhao, L. Fu, Y. Gu, L. Qu, C. Zhu, J. Zhu and Y. Lin, Chem. Rev., 2023, 479, 215002.
19. P. Zuo, L. Jiang, X. Li, P. Ran, B. Li, A. Song, M. Tian, T. Ma, B. Guo, L. Qu and Y. Lu, Nanoscale, 2019, 11, 485–494.
20. G. Song, S. Cong and Z. Zhao, Chem. Sci., 2022, 13, 1210–1224.
21. Y. Liu, Z. Gao, M. Chen, Y. Tan and F. Chen, Adv. Funct. Mater., 2018, 28, 1805710.
22. X. Tang, X. Fan, J. Zhou, S. Wang, M. Li, X. Hou, K. Jiang, Z. Ni, B. Zhao, Q. Hao and T. Qiu, Nano Lett., 2023, 23, 7037–7045.
23. Z. Zheng, S. Cong, W. Gong, J. Xuan, G. Li, W. Lu, F. Geng and Z. Zhao, Nat. Commun., 2017, 8, 1993.
24. Q. Lv, J. Tan, Z. Wang, L. Yu, B. Liu, J. Lin, J. Li, Z. Huang, F. Kang and R. Lv, Adv. Funct. Mater., 2022, 32, 2200273.
25. Y. Zhang, Z. Shi, H. Cui, Q. Xia, F. Liu, Z. Wang, J. Wang, H. Fan, C. Shu, B. Chen, H. Li, Z. Lai, Z. Luo, W. Zheng, L. Wang and Z. Huang, Nano Lett., 2024, 24, 14293–14301.
26. L. Tao, Z. Li, K. Chen, Y. Zhou, H. Li, X. Wang, R. Zhan, X. Hou, Y. Zhao, J. Xu, T. Qiu, X. Wan and J. Xu, Cell Rep. Phys. Sci., 2021, 2, 100526.
27. G. Song, W. Gong, S. Cong and Z. Zhao, Angew. Chem., 2021, 60, 5505–5511.
28. G. Xu, R. Dong, D. Gu, H. Tian, L. Xiong, Z. Wang, W. Wang, Y. Shao, W. Li, G. Li, X. Zheng, Y. Yu, Y. Feng, Y. Dong, G. Zhong, B. Zhang, W. Li, L. Wei, C. Yang and M. Chen, J. Phys. Chem. Lett., 2022, 13, 1453–1463.
29. L. Jiang, J. Zhang, H. Li, C. He, R. Yuan, Y. Zhao, J. Huang and X. Yang, Chem. Eng. J., 2023, 461, 142102.
30. X. Meng, J. Yu, W. Shi, L. Qiu, K. Qiu, A. Li, L. Zhen, Y. Wang, J. Wu, J. Lin, X. Wang and L. Guo, Angew. Chem., 2024, 136, e202407597.
31. Y. Zhou, Q. Gu, T. Qiu, X. He, J. Chen, R. Qi, R. Huang, T. Zheng and Y. Tian, Angew. Chem., 2021, 133, 26464–26471.
32. S. Xie, K. Lai, C. Gu, T. Jiang, L. Zhou, X. Zheng, X. Shen, J. Han and J. Zhou, Mater. Today Nano, 2022, 18, 100179.
33. H. Luo, B. Wang, F. Wang, J. Yang, F. Wu, Y. Ning, Y. Zhou, D. Wang, H. Liu and S. Dou, ACS Nano, 2020, 14, 7328–7337.
34. R. Li, S. Chen, X. Zhang, F. Zeng, X. Song, J. Yin and C. Jiang, Sci. China Mater., 2024, 67, 2985–2994.
35. G. Kumari, K. Jayaramulu, T. Maji and C. Narayana, J. Phys. Chem., 2013, 117, 11006–11012.
36. Y. Ma, D. Xiong, Y. Meng, Y. Lu, P. Duan, B. Chen, Y. Yang, X. Chen, L. Mao, X. Wu and L. Yang, Sci. China Mater., 2025, 1–11.
37. L. Colombo, V. VolovSek, K. Furik and J. Durig, J. Raman Spectrosc., 1990, 21, 169–175.
38. X. Li, F. Zhang, M. Zhang, X. Zhou and H. Zhang, ACS Omega, 2021, 6, 35548–35558.
39. R. AlGaashani, S. Radiman, A. Daud, N. Tabet and Y. Al-Douri, Ceram. Int., 2013, 39, 2283–2292.
40. G. Silversmit, D. Depla, H. Poelman, G. Marin and R. Gryse, J. Electron Spectrosc. Relat. Phenom., 2004, 135, 167–175.
41. S. K. Das, S. K. Nandi, C. V. Marquez, A. Rua, M. Uenuma, E. Puyoo, S. K. Nath, D. Albertini, N. Baboux, T. Lu, Y. Liu, T. Haeger, R. Heiderhoff, T. Riedl, T. Ratcliffe and R. G. Elliman, Adv. Mater., 2023, 35, 2208477.
42. M. Tarannum and S. Egusa, J. Phys. Chem. C, 2024, 128, 4125–4223.
43. L. Zhang, J. Ran, S. Z. Qiao and M. Jaroniec, Chem. Soc. Rev., 2019, 48, 5184–5206.

---